Random Graphical Model of Microbiome Interactions in Related Environments

• DOI: 10.1007/s13253-024-00638-6
• Publication Date: 2024-06-20T15:02:01Z
• Authors: Veronica Vinciotti; Ernst C. Wit; Francisco Richter
• Keywords: Methodology (stat.ME); FOS: Computer and information sciences; 0301 basic medicine; 03 medical and health sciences; Microbiome; Graphical models; Random graph model; Bayesian inference; Statistics - Methodology

Abstract The microbiome constitutes a complex microbial ecology of interacting components that regulates important pathways in the host. Most communities at various body sites tend to share common substructures interactions, while also showing diversity related needs local environment. aim this paper is develop method for inferring both core and differences such microbiota systems. approach combines two elements: (i) random graph model generating networks across environments, capturing potential relatedness structural level, with (ii) Gaussian copula graphical inference environment-specific from multivariate data. We propose Bayesian joint systems metagenomic data number sites. analysis human shows how proposed able capture varying levels similarity different supported by their taxonomical classification. Beyond stable core, inferred show interesting between sites, as well interpretable relationships classes microbes.